The well-known advantages of a polymer having chains containing directly linked polycyclic repeating units free of unsaturation, have driven those skilled in the art to search for a processable "addition polymer" of one or more multi-ringed monoolefinically unsaturated cycloolefin monomers such as norbornene, bicyclo[2.2.1]hept-2-ene or "NB" for brevity, and substituted embodiments thereof, such as ethylidenenorbornene or decylnorbornene, and particularly those monomers of NB having at least one substituent in the 5- (and/or 6-) positions. The foregoing monomers are collectively referred to herein as "norbornene-type" or "norbornene-functional" or "NB-type" or "NB-functional" monomers, for convenience, recognizing that, just as in NB, or substituted NB, each NB-type polymer is characterized by containing a repeating unit resulting from an addition polymerized derivative of bicyclo[2.2.1]hept-2-ene. A first NB-type or NB-functional monomer may be polymerized by coordination polymerization to form (i) an addition homopolymer; or, (ii) with a second NB-type or NB-functional monomer, either one (first or second) of which is present in a major molar proportion relative to the other, to form an addition NB-type copolymer; or, (iii) with a second monomer which is not an NB-type monomer, present in a minor molar proportion relative to the first, to form an addition copolymer with plural repeating units of at least one NB-type or NB-functional monomer.
Polynorbornene or "poly(bicyclo[2.2.1]hept-2-ene)" or polyNB for brevity, was originally produced a long time ago (U.S. Pat. No. 2,721,189). However this original material was found to contain two types of polymers, one brittle, the other thermoformable and `drawable`. The brittle polymer was later found to be a low molecular weight (`mol wt`) saturated polymer which was termed an addition type polymer; and, the thermoformable polymer was shown to be formed by ring opening metathesis polymerization (`ROMP`). A ROMP polymer has a different structure compared with that of the addition polymer in that (i) the ROMP polymer of one or more NB-type monomers, contains a repeat unit with one less cyclic unit than did the starting monomer, and, (ii) these are linked together in an unsaturated backbone characteristic of a ROMP polymer and is shown below. ##STR1##
It will now be evident that, despite being formed from the same monomer, an addition-polymerized polyNB is clearly distinguishable over a ROMP polymer. Because of the different (addition) mechanism, the repeating unit of the former has no backbone C.dbd.C unsaturation as shown below: ##STR2##
The difference in structures of ROMP and addition polymers of NB-functional monomers is evidenced in their properties, e.g., thermal properties. The addition type polymer of NB has a high T.sub.g of about 370.degree. C. The unsaturated ROMP polymer of NB exhibits a T.sub.g of about 35.degree. C., and exhibits poor thermal stability at high temperature above 200.degree. C. because of its high degree of C.dbd.C unsaturation.
Some time later, reaction conditions were optimized so as to enable one to choose, and selectively make, either the low mol wt addition polymer, or the ROMP polymer. In U.S. Pat. No. 3,330,815, the disclosure taught that only the addition polymer was synthesized with TiCl.sub.4 /Et.sub.2 AlCl or Pd(C.sub.6 H.sub.5 CN).sub.2 Cl.sub.2, under particular conditions, except that the polymers produced were only those in the mol wt range from 500 to 750 in which range they were too brittle for any practical application.
Addition polymers of norbornene have been shown to be produced with "zirconocene type" catalysts such as those taught by Kaminsky et al, and others, all well known to those skilled in the art. These polymers have been found to be a highly crystalline form of a "norbornene-addition polymer", that is, an addition polymer of a NB-functional monomer, which is totally insoluble, and reportedly does not melt until it decomposes at .apprxeq.600.degree. C. (under vacuum to avoid oxidation). It is therefore unprocessable (W. Kaminsky et al., J. Mol. Cat. 74, (1992), 109; W. Kaminsky et. al. Makromol. Chem, Macromol. Symp., 47, (1991) 83; and W. Kaminsky, Shokubai, 33, (1991) 536.). An added distinguishing characteristic of the zirconocene catalyst system is that it catalyzes the copolymerization of ethylene and norbornene. In such copolymers, the amount of NB incorporated into the ethylene/NB copolymer can be varied from high to low (W. Kaminsky et. al. Polym. Bull., 1993, 31, 175).
The polymer formed with a zirconocene catalyst can incorporate ethylene (or compounds containing ethylenic unsaturation at a terminal end thereof) in its backbone, randomly, whether in runs of a multiplicity of repeating units, or even a single unit. It should also be noted that the ionic metallocene catalysts, such as zirconocenes and hafnocenes, use metals from Group IVB as the cation with a compatible weakly coordinating anion. These catalysts are entirely distinct from the catalysts used in this invention.
Research has continued toward the production of a melt-processable addition polymer of a NB-type monomer, and is the subject of an on-going effort. By "melt-processable" it is meant that the polymer is adequately flowable to be thermoformed in a temperature window above its T.sub.g but below its decomposition temperature. To date, there has been no disclosure of how to solve the many problems inherent in the production of a heat-resistant, yet thermoformable mid processable polymer of a NB-functional monomer which polymer can be extruded, injection molded, blow molded, and the like, using conventional equipment.
To date, we know of no practical or reliable method for commercially producing an amorphous NB addition polymer with controlled mol wt. Polymers formed with too low a mol wt are of limited utility in thermoformed articles. Polymers with too high a mol wt can only be cast from solution and in some cases are completely insoluble and difficult to thermoform. The goal has been to produce an addition polymer having a mol wt M.sub.w in the range of 50,000 to 500,000, using only one or more NB-functional monomers, in a reliably controlled manner. The only method available to produce such a polymer has been through premature deactivation of the catalyst systems which produce amorphous polymers of NB, the homopolymers having mol wts in the millions. Predictably, this method of mol wt control leads to low catalyst productivity and requires the use of high catalyst levels when the mol wt M.sub.w is to be in the range from about 150,000-350,000. Since the problem of forming a processable NB-type polymer was never solved, the second, equally serious problem of obtaining a useful or practical level of conversion was never addressed.
A few years ago the reactivity of cationic, weakly ligated, transition metal compounds was studied in the polymerization of olefins and strained ring compounds, (A. Sen, T. Lai and R. Thomas, J. of Organometal. Chemistry 358 (1988) 567-568, C. Mehler and W. Risse, Makromol. Chem., Rapid Commun. 12, 255-259 (1991)). Pd complexes incorporating the weakly ligating CH.sub.3 CN (acetonitrile) ligand in combination with a weakly coordinating counteranion could only be used with aggressive solvents such as acetonitrile or nitromethane. When Sen et al used the complexes to polymerize NB, a high yield of a homopolymer which was insoluble in CHCl.sub.3, CH.sub.2 Cl.sub.2 and C.sub.6 H.sub.6, was obtained.
The identical experimental procedure, with the same catalyst and reactants, when practiced by Risse et al used one-half the molar amount of each component. Risse et al reported the synthesis of a polyNB homopolymer which had a mol wt M.sub.n of 24,000. In other runs, using different ratios of NB to Pd.sup.2+ -compound, polyNBs having mol wts M.sub.n of 38,000 and 70,000 respectively with narrow dispersities M.sub.w /M.sub.n in the range from 1.36 to 1.45, and viscosities in the range from 0.22 to 0.45 dL/g were made. A homopolymer which had a viscosity of 1.1 was synthesized, which upon extrapolation from the mol wt data given for the prior runs, indicates the M.sub.w was over 10.sup.6. See Mehler and Risse Makromol. Chem., Rapid Commun. 12, 255-9 (1991), experimental section at the bottom of page 258 and the GPC data in Table 1 on pg 256. The polymers were soluble in 1,2-dichlorobenzene in which Risse et. al. measured mol wts by GPC (gel permeation chromatography) and viscometry, as did Maezawa et al in EP 445,755A, discussed below.
Maezawa et al disclosed the production of high mol wt NB polymers with a two-component catalyst system. The disclosure states that the polymer is preferably formed in the molecular weight range from 10.sup.5 to 10.sup.7. The manner of obtaining the desired mol wt is shown to be by terminating the polymerization reaction after a predetermined period. Such termination is effected by decomposing the catalyst with an external terminating agent such as acidified methanol, which is added to the reaction to stop the polymerization. There is no internal control of the mol wt within a predetermined range by an agent that does not deactivate the catalyst.
Specifically, three known methods of controlling the mol wt are suggested: (i) varying the amount of the transition metal compound used; (ii) varying the polymerization temperature; and (iii) using hydrogen as a chain transfer agent "CTA" (see page 9, lines 20-23 of the '755A disclosure) as suggested by Schnecko, Caspary and Degler in "Copolymers of Ethylene with Bicyclic Dienes" Die Angewandte Makromolekulare Chemic, 20 (1971 ) 141-152 (Nr.283). Despite the foregoing suggestions, there is no indication in '755A that any of them was effective, as is readily concluded from the illustrative examples in the specification. As stated in their illustrative Example 1 in which the catalyst included a combination of nickel bisacetylacetonate Ni(acac).sub.2 and methaluminoxane ("MAO"), a polyNB having M.sub.w =2.22.times.10.sup.6 (by GPC) was formed. As shown in Table 1 of the '755A reference, only Exs. 5, 6 and 7, in which the (triphenylphosphine)Ni-containing catalysts were used, made homopolymers with M.sub.w =234,000; 646,000; and 577,000 respectively. These nickel catalysts with a triphenylphosphine ligand, are shown to have relatively lower productivity than the biscyclooctadienylnickel (Ex 3) and biscyclopentadienylnickel (Ex 4) which were also used.
One is therefore led to conclude that only those Ni-based catalysts which have substantially lower productivity than Ni(acac).sub.2 with a MAO catalyst system would effectively decrease the mol wt of the homopolymer produced. There is no suggestion that any of the polymers disclosed in the '755A reference are likely to be melt-processable. A conclusion that they are not melt-processable is supported by the evidence that all the polymers made by Maezawa et al were cast from solution.
A key aspect of the '755A disclosure was that the catalyst system disclosed was a combination of at least two components, namely, a transition metal complex, and a methaluminoxane cocatalyst. Maezawa et al used this multi-component catalyst system to produce the high mol wt polymers in the range above 5.times.10.sup.5. It was critical that the transition metal component in the complex be from Groups VB, VIB, VIIB, and VIII, and that it be paired with the methaluminoxane cocatalyst in order to produce polymer in a reasonable yield. The criticality of the cocatalyst was confirmed by illustrative examples of transition metal compounds which were generally catalytically effective only so long as methaluminoxane was the cocatalyst (Comparative Examples 3, and 4). The experimental evidence indicated that attaining a high productivity catalyst system was limited to specific nickel complexes in combination with MAO as the activator. All the illustrative examples having been run in toluene, it is evident that they were unaware that a polar solvent such as a halohydrocarbon, and the like, might improve productivity.
It is evident that the results obtained with the '755A catalyst/cocatalyst system are different from those with a Group VIII metal catalyst in which the metal is weakly ligated to displaceable ligands and a portion of a ligand generates a .sigma.-bond. Whether the .sigma.-bond-generating ligand has an allyl group or a canonical form thereof, the allyl metal linkage provides the initial metal-C .sigma.-bond into which successive NB-type moieties are inserted to form a polymer chain. This insertion reaction is well known in the analogous propagation of ethylene in Ziegler Natta catalysis described in detail in the text Comprehensive Organometallic Chemistry edited by Geoffrey Wilkinson et al, in a chapter titled "Ziegler-Natta Catalysis" by Gavens et al, 1982, pg 484 et seq. Allyl-Ni cationic complexes have been synthesized for the polymerization of butadiene, but an allyl-Ni-cyclooctadiene ("allyl-Ni COD") cation complex was reported not to be catalytically active (see text, The Organic Chemistry of Nickel P. W. Jolly and G. Wilke, Vol I Academic Press New York, 1974 pg 352).
On the other hand, it has long been recognized that cationic nickel compounds are active catalysts for the polymerization of butadiene (R. Taube, et al Makromol. Chem., Macromol. Symp. 66, (1993) 245; L. Porri, G. Natta, M. C. Gallazzi, J. of Polymer Sci. Pt C. 16 (1967) 2525). Taube et al state "The chain growth proceeds by the insertion of butadiene into the allyl nickel bond always with formation of the new butenyl group in the `anti` configuration (anti insertion)." The coordination of an allyl type ligand to the nickel is maintained continuously throughout the butadiene polymerization. This mechanism is clearly distinguishable from the insertion mechanism of a NB-functional monomer in which insertion of only the very first monomer molecule occurs at an allyl type ligated metal center.
Allylnickelhalides alone (no Lewis acid cocatalyst) have been used to produce polyNB, however the molecular weights of the NB polymer produced in these studies were actually low; e.g. 1000 to 1500 mol wt. (L. Porri, G. Natta, M. C. Gallazzi Chim. Ind. (Milan), 46 (1964),428). It had been thought that the low yields and the low mol wts of the polyNB were due to deactivation of the catalysts.
Still more recently, in a lone example of the use of a nickel catalyst as a transition metal equivalent to zirconium, Okamoto et al disclosed the production of high mol wt norbornene polymer with a three component catalyst system in example 117 on page 46 of EP 504,418A. The three-component catalyst was made in situ by combining triisobutylaluminum; dimethylanilinium tetrakis(pentafluorophenyl)borate; and, Ni(acac).sub.2 in toluene. The polymer recovered had a M.sub.w =1.21.times.10.sup.6 and a mol wt distribution of 2.37. Though essentially the entire specification is directed to the copolymerization of cycloolefins with .alpha.-olefins using zirconium-containing catalysts, Okamoto et al did not react norbornene and .alpha.-olefin with a nickel catalyst. Nowhere in the '418A specification is there a teaching that the use of an .alpha.-olefinic CTA will control molecular weight. There is no teaching of a polymer with a terminal olefinic end-group. Nor is there any teaching that an .alpha.-olefin would do anything but copolymerize.
The failure to recognize that an .alpha.-olefin might function as a CTA, with or without the presence of an alkylaluminum cocatalyst, was understandable since there existed a large body of work related to the copolymerization of cycloolefins with .alpha.-olefins, and in none of such polymerizations was there any disclosure that the .alpha.-olefin might function as an effective CTA. Further, the great reactivity of ethylene or propylene buttressed an expectation that copolymerization, not chain transfer, is the logical and expected result.
Since practical consideration relating to melt-processing cycloolefin addition polymers produced herein, dictate that their mol wt be controlled within one order of magnitude, e.g., in the range from 50,000 to 500,000, it is evident that the '755A invention was unable to provide either a solution to the problem, or even an enabling disclosure to solve it. They do not suggest they can reliably make a reproducible polymer in the defined mol wt range. They suggest the use of hydrogen as CTA, and provided no reason to explore using another, least of all a CTA with a terminal non-styrenic, non-vinyl ether double bond. Moreover, there is no disclosure of a polymer with a terminal end-group derived from a compound having terminal unsaturation.
Neither is there any basis for estimating the effect of an .alpha.-olefin as a CTA in an insertion reaction, particularly insofar as the .alpha.-olefin is effective to tailor the mol wt of the growing polymer chains in an addition polymerization, irrespective of whether a multicomponent Group VIII catalytic system is used in a complex catalyst of the type taught by Maezawa.
An acyclic olefin. e.g., 1-hexene, is known to be an effective CTA in the ROMP of cyclic olefins, to reduce mol wt via a cross-metathesis mechanism. ROMP involves a metal carbene (or metal alkylidene) active center which interacts with the cyclic olefin monomer to afford a metallocycloalkane intermediate. A repeating unit contains a C.dbd.C double bond for every C.dbd.C double bond in the monomer. How effectively the acyclic olefin reduces the mol wt of the copolymer formed depends on the structure of the olefin and on the catalyst system (K. J. Ivin, Olefin Metathesis, Academic Press, 1983). In contrast. addition (or vinyl type) polymerization of olefins and diolefins involves the insertion of the monomer into a metal-carbon a-bond, as in Ni--C, or Pd--C. Despite the many disclosures relating to the formation of copolymers of NB-type monomers, and the well-known fact that an olefin is an effective chain transfer agent in a ROMP polymerization, it will now be evident why the difference in the mechanisms of chain termination failed to suggest the use of an olefin as a chain transfer agent in the copolymerization taught herein.
Chain transfer via .beta.-hydride elimination has been previously described. See, for example, Polypropylene and other Polyolefins Polymerization and Characterization by Ser van der Ven, Studies in Polymer Science 7, Elsevier Amsterdam, etc. 1990, Chapter 1 POLYPROPYLENE; CATALYSTS AND POLYMERIZATION ASPECTS by Brian L. Goodall, and Section 1.6 thereof titled "The Effect of Catalyst and Process Variables on the Molecular Weight and its Distribution ("Chain Transfer"), and particularly Section 1.6.3 On The Mechanism of Chain Transfer, pg 82-83. In typical .alpha.-olefin polymerization hydrogen is introduced to control molecular weight. There is no teaching that the introduction of a second type of olefin will result in control of molecular weight or will selectively terminate a polymer chain with a well-defined olefinic end-group.
In typical .alpha.-olefin polymerizations, it is recognized that the known mechanism of ".beta.-hydride elimination" can provide a double bond near the terminal end of the polymer chain. In this mechanism which modulates the mol wt of olefinic polymers, a metal bonded to a hydrocarbyl radical with hydrogens on the carbon .beta. to the metal, can undergo a reaction where the .beta.-hydrogen is abstracted to the metal, leaving an olefinic group. This results in an unsaturated polymer chain and the metal hydride. In general, the rate of .beta.-hydride elimination vs. the polymerization rate, controls the molecular weight of the polymer. For most polymerization catalyst systems, the proclivity of the catalyst system toward .beta.-hydride elimination must be extensively researched and is not predictable. The polymer mol wt depends upon a host of process variables: the choice of monomer or monomers, the presence of or absence of hydrogen, the ligand environment around the transition metal, the presence of additional donor ligands, type of catalyst (homogeneous or heterogeneous), presence or absence of a cocatalyst (and choice thereof), and polymerization medium (bulk; solution, slurry, gas phase), inter alia. It will be clear from the above and is well documented in the literature that the resulting poly(.alpha.-olefin) contains a mixture of end-groups both saturated and unsaturated.
The factors that influence .beta.-hydrogen elimination in the case of Group VIII metal catalysts are also unpredictable, for example, nickel catalysts have been used in the polymerization of ethylene. Depending upon the Ni catalyst chosen, it is possible to generate exclusively the dimer (1-butene), higher olefins (oligomers), or high mol wt polyethylene. Homogeneous Ni catalysts for the polymerization of ethylene to high mol wt polyethylene have been described by Klabunde et al. (U. Klabunde et al., J. Polym. Sci., Polym. Chem., 25 p 1989 (1987)) and Ostoja Starzewski (P. W. Jolly and G. Wilke, Vol 2, supra) where the polymer mol wt is controlled by the ligand environment around the nickel and the choice of reaction medium. The polymerization of ethylene has been reported to occur in the presence of a variety of different nickel containing Ziegler catalysts and single-component nickel catalysts, while other nickel catalysts give only dimers (see P. W. Jolly and G. Wilke, Vol 2, supra). Shell Oil Co. uses a nickel-catalyzed oligomerization of ethylene to manufacture linear .alpha.-olefins on a large scale (see G. W. Parshall and S. D. Ittel, Homogeneous Catalysis: The Applications and Chemistry of Catalysis by Soluble Transition Metal Complexes, John Wiley and Sons, 1992).
The mechanism by which an .alpha.-olefin affects both initiation and propagation rate in a different polymerization system, namely the cobalt-catalyzed polymerization of butadiene (to butadiene rubber) was known, as stated by Goodall supra, on pg 83, but the rate at which the reaction occurs, and the amount of butadiene which is incorporated in the rubber chains is not predictable, necessitating the presence m the reactor, of a major molar amount of .alpha.-olefin relative to the butadiene. From the foregoing considerations there is no basis in the art to predict the effect of an .alpha.-olefin on the polymerization of a NB-functional monomer.
It is noted that nickel catalysts have been used in the polymerization of butadiene where the solvent is neither a chlorohydrocarbon nor an aromatic solvent such as toluene or xylene. With some catalysts, the microstructure of the polymer is a function of its mol wt. With others, unsaturated hydrocarbons such as acetylenes and allenes retard initiation and propagation and enhance chain transfer, but do not affect microstructure (see Encyclopedia of Polymer Science and Engineering Second Edition, Vol 2, pg 537; John Wiley and Sons, 1985).
However monoolefins were reported to have no effect on the polymerization of butadiene, at least when the amount added is relatively small (see R. Sakata, J. Hosono, A. Onishi and K. Ueda, Makromol. Chem., 139 (1970) 73). Still other nickel catalysts (with different ligand environments) give only (cyclic) dimers and trimers, such as "COD" and cyclododecatriene ("CDT").
It should be noted that the structure of the Ni-cyclodiolefin complex has been investigated in the interest of exploring numerous transition metal complexes with weakly ligated compounds in combination with a counteranion. Such a study was published by K. Kempe and J. Sieler in Zeitschrift fur Kristallographie 201, 287-289 (1992) who did not suggest it would have catalytic activity. Also known are compounds related to (.pi.-C.sub.3 H.sub.5 NiCl).sub.2.TiCl.sub.4, which compounds are formed by reacting .pi.-allylnickel halides with strong Lewis acids (e.g., TiCl.sub.4, AlBr.sub.3), and these are used for the polymerization of butadiene and the dimerization of olefins. There was no logical reason from known facts about nickel catalysts which would suggest the use of the known metal complex as a particularly effective catalyst for NB-functional monomers.
There is a need to control the T.sub.g of NB-addition polymers. The effect of an alkyl substituent on the T.sub.g of a copolymer was disclosed in an article titled "Synthesis and characterization of poly(5-alkyl-2-norbornene)s by cationic polymerization. Effect of alkyl substituent length on monomer reactivity, polymer structure and thermal properties" by T. Sagane et al, Macromol. Chem. 4, 37-52 (1993). The longer the sidechain, the lower the T.sub.g of the polymer. However, the copolymers were made with a AlEtCl.sub.2 /tert-butyl chloride catalyst system, and the mol wt M.sub.w of the longest chain made was less than 2500. There was no suggestion that any other complex metal system, or any other catalyst system might yield higher mol wts.
In view of the foregoing discussion, the prior art has not described or contemplated NB-type addition polymers having a single olefinic group located at a terminal end thereof. Nor has the prior art described or contemplated a method of controlling the molecular weight of an addition polymerized NB-type polymer in the presence of a chain transfer agent having a terminal double bond. Moreover, there is no teaching that the introduction of a selected .alpha.-olefin CTA into the reaction medium will selectively terminate a NB-type addition polymer chain with a well-defined olefinic end-group. Additionally, the prior art does not address the effect of alkyl substituent length for the control of T.sub.g of NB-type polymers over 2,500 M.sub.w.